Hamed Hassani

papers

778

total citations

papers (19)

Linear Convergence in Federated Learning: Tackling Client Heterogeneity and Sparse Gradients

NEURIPS 2021arXiv

182

citations

A Theory of Non-Linear Feature Learning with One Gradient Step in Two-Layer Neural Networks

ICML 2024arXiv

citations

T-Cal: An Optimal Test for the Calibration of Predictive Models

ICML 2024arXiv

citations

Decision Theoretic Foundations for Conformal Prediction: Optimal Uncertainty Quantification for Risk-Averse Agents

ICML 2025arXiv

citations

Conformal Prediction with Learned Features

ICML 2024arXiv

citations

Adversarial Reasoning at Jailbreaking Time

ICML 2025arXiv

citations

Adversarial Training Should Be Cast as a Non-Zero-Sum Game

ICLR 2024arXiv

citations

Sinkhorn Natural Gradient for Generative Models

NEURIPS 2020arXiv

citations

Provable Multi-Task Representation Learning by Two-Layer ReLU Neural Networks

ICML 2024arXiv

citations

Sinkhorn Barycenter via Functional Gradient Descent

NEURIPS 2020arXiv

citations

Collaborative Linear Bandits with Adversarial Agents: Near-Optimal Regret Bounds

NEURIPS 2022arXiv

citations

Submodular Meta-Learning

NEURIPS 2020arXiv

citations

Collaborative Learning of Discrete Distributions under Heterogeneity and Communication Constraints

NEURIPS 2022arXiv

citations

Conformal Information Pursuit for Interactively Guiding Large Language Models

NEURIPS 2025arXiv

citations

Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth

ICML 2024arXiv

citations

Hamed Hassani

papers (19)

Linear Convergence in Federated Learning: Tackling Client Heterogeneity and Sparse Gradients

Model-Based Domain Generalization

FedAvg with Fine Tuning: Local Updates Lead to Representation Learning

Probable Domain Generalization via Quantile Risk Minimization

Adversarial Robustness with Semi-Infinite Constrained Learning

A Theory of Non-Linear Feature Learning with One Gradient Step in Two-Layer Neural Networks

T-Cal: An Optimal Test for the Calibration of Predictive Models

Decision Theoretic Foundations for Conformal Prediction: Optimal Uncertainty Quantification for Risk-Averse Agents

Conformal Prediction with Learned Features

Adversarial Reasoning at Jailbreaking Time

Adversarial Training Should Be Cast as a Non-Zero-Sum Game

Sinkhorn Natural Gradient for Generative Models

Provable Multi-Task Representation Learning by Two-Layer ReLU Neural Networks

Sinkhorn Barycenter via Functional Gradient Descent

Collaborative Linear Bandits with Adversarial Agents: Near-Optimal Regret Bounds

Submodular Meta-Learning

Collaborative Learning of Discrete Distributions under Heterogeneity and Communication Constraints

Conformal Information Pursuit for Interactively Guiding Large Language Models

Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth

papers (19)

Linear Convergence in Federated Learning: Tackling Client Heterogeneity and Sparse Gradients

Model-Based Domain Generalization

FedAvg with Fine Tuning: Local Updates Lead to Representation Learning

Probable Domain Generalization via Quantile Risk Minimization

Adversarial Robustness with Semi-Infinite Constrained Learning

A Theory of Non-Linear Feature Learning with One Gradient Step in Two-Layer Neural Networks

T-Cal: An Optimal Test for the Calibration of Predictive Models

Decision Theoretic Foundations for Conformal Prediction: Optimal Uncertainty Quantification for Risk-Averse Agents

Conformal Prediction with Learned Features

Adversarial Reasoning at Jailbreaking Time

Adversarial Training Should Be Cast as a Non-Zero-Sum Game

Sinkhorn Natural Gradient for Generative Models

Provable Multi-Task Representation Learning by Two-Layer ReLU Neural Networks

Sinkhorn Barycenter via Functional Gradient Descent

Collaborative Linear Bandits with Adversarial Agents: Near-Optimal Regret Bounds

Submodular Meta-Learning

Collaborative Learning of Discrete Distributions under Heterogeneity and Communication Constraints

Conformal Information Pursuit for Interactively Guiding Large Language Models

Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth